This paper gives an insight into next generation maintenance systems, by examining the limitations of present maintenance systems and forecasting future needs based on business challenges and technological advancements.
Maintenance Past and Present
Maintenance is done to preserve the functioning of our equipment and as such it should be continued to be done for as long as it is economical.
Maintenance as a discipline has changed over the years. This can be attributed to technological developments; more complex equipment designs. These complex designs have meant changes to maintenance techniques as a result of greater expectations from management. These changes have also changed maintenance requirements in terms of how and who carries out maintenance.
What is referred to as first generation, maintenance types were fix it when it broke, run to failure. Downtime was not an issue, as there was a low level of automation and equipment was reliable and highly maintainable. This was as a result of simple designs which in most cases over designed.
The second generation was characterised by increased pressure in the demand for goods and services. This was associated with manpower shortages, rising cost and high levels of inflation as this period of time was, during and post Second World War till about the late 1970's.
In order to successful deal with this, mechanisation was on the increase. This meant the use of more complex equipment and a change to the manner in which maintenance was performed as downtime now became an issue. Maintenance task moved from run to failure to preventive forms of maintenance. Maintenance costs were on the rise, to deal with this maintenance systems also changed in regards to planning, scheduling and control. Safety standards, product quality have also changed. The end result could be summarised with the introduction of maintenance management systems
What is termed third generation, spans from about the mid seventies till present. This has been characterised by greater expectations in equipment reliability, availability and cost.
This has been as a result of major engineering disasters such as Piper Alpha, Bhopal and Chernobyl. These events have changed have resulted in the way maintenance is viewed. This is because of increased awareness in terms of; safety quality, environment and changes in statutory requirements. This is in addition to considering life cycle costing. Maintenance is being gradually being recognised as a strategic issue and some have referred to it as the 'last goldmine'. It is in view of this that maintenance techniques like condition monitoring and maintenance methodologies such as Reliability Centred Maintenance (RCM) and Total Productivity Management (TPM) have been developed and are being implemented.
Maintenance can be defined as the management of, control, execution and quality of those activities which will ensure that optimum levels of availability and overall performance of plant are achieved in order to meet business objectives (UK Department of Trade and Industry - DTI - 1991)
It can be said that from the above definition that maintenance is concerned with the entire life cycle of an asset not just a part of it.
Computer Maintenance Management System (CMMS)
The CMMS is used to coordinate and control the various subsystems. This is in light of the evolution in maintenance which has been discussed earlier. This has resulted in increasing amounts of data and shorter time to make decisions. Computer have been introduced to help processes this data with the objective of reducing the time to take decisions.
The CMMS is now a key component in maintenance and offers various levels of support in the organisation hierarchy which are as follows: [6]
It supports condition based monitoring of machines and assets; to offer insight into wear and imminent failures;
It can track the movement of spare parts and requisition replacements when necessary;
It allows operators to report faults faster, thus enabling maintenance staff to respond to problems more quickly;
It can facilitate improvements in communication between operations and maintenance personnel, and is influential in improving the consistency of information passed between these two departments;
It provides maintenance planners with historical information necessary for developing PM schedules;
It provides maintenance managers with information in a form that allows for more effective control of their department's activities;
It offers accountants information on machines to enable capital expenditure decisions to be taken; and
It affords senior management a crucial insight into the state of asset care within their organisation.
A CMMS is ideally a means to achieving world class maintenance, as it provides a platform for decision analysis thereby acting as a guide to management [7]. The benefits of using this system include;[8]
Resource control - tighter control of resources ;
Cost management - better cost management and audibility;
Scheduling - ability to schedule complex, fast moving workloads; integration with other business systems; and
Reduction of breakdowns - improved reliability of physical assets through the application of an effective maintenance programme.
In spite of the outstanding benefits of using the CMMS, it has failed in regards to achieving the fundamental function of maintenance which is to reduce the number of breakdowns.
It can then be said that near zero downtime will be the driving force in the next generation of maintenance systems.
According to the IMS the unmet needs of maintenance systems today are:
Machine Intelligence: Intelligent monitoring, predict and prevent, and compensation, reconfiguration for sustainability (self-maintenance).
Operations Intelligence: Prioritize, optimize, and responsive maintenance scheduling for reconfiguration needs.
Synchronization Intelligence: Autonomous Information Flow from Market Demand to Factory Asset Utilization.
Labib 2002 breaks these unmet needs to the following research questions;
Adapting PM schedules to cope dynamically with shop floor reality
Feedback information gather in maintenance to the designers
Linking maintenance policies to the corporate strategy
Synchronising production scheduling based on maintenance performance
Labib further break it down into areas of further discussion
Self reconfiguration
Self maintaining/ healing
Data - information- decisions
Offline-online- real-time
Learn and grow
Diagnostics to prognostics
This report examines the available technology and ongoing research in these areas. The technology here can be used to meet the unmet needs in maintenance systems and in the design of the next generation maintenance systems.
Self maintaining
Self maintaining machines are able to maintain its function for awhile even though a fault happens.
The following are the requirements for a self maintaining machine (ref)
Monitoring capability
Fault judging capability
Diagnosing capability
Repair planning capability
Repair executing capability
These requirements are supported by Wang et al 2010, stating that a self maintaining system would have to consist of a detection system, a decision system and a maintenance system. Linking the requirements made by Labib 2006 and Wang et al 2010
The detection system would embody the monitoring, fault judging and diagnosing capability, the decision system would embody the repair planning capability and the maintenance system the repair executing capability.
The technologies for the systems would be said to cover fault detection, health management (both of which are achieved through condition monitoring), decision support system (which involves data processing to give meaningful information) and artificial technology.
Lee et al 2006 adds that another approach to self maintaining systems, which is having a self trigger function. This function sends a request for maintenance with detailed requirements for what needs to be done to a maintenance team. 'This integration of machine, maintenance schedule, dispatch system and inventory management system will minimize maintenance costs to the greatest'.
It must be said that self maintenance is what Lee et 2006 refer to as 'functional maintenance' which is the capability of a machine to still perform its indeed function even after failure or degradation have occurred, by channelling resources to where it is needed to ensure that it continues to perform its intend function. This is unlike 'physical maintenance' (traditional repairs) which might involve physically changing a faulty or failure component in the system.
An example of a self maintaining system carrying out 'functional maintenance' is the self resilient robot developed by Alexander Gloye-Förster and his team. The robot was used in the 2004 RoboCup. 'The robots can autonomously recover from certain types of unexpected damage, through adaptive self-models derived from actuation-sensation relationships, used to generate forward locomotion'.
For a system to carry out self maintenance it requires a level of machine intelligence. The machine must be capable enough of carrying out functional maintenance.
Self Healing
This is a property of a material to restore itself to its original set of properties.
This is a novel alternative to damage tolerant design and removes the need to perform temporary repairs to damaged structures (Williams et al 2002). It is not limited to this but also includes mitigating the damages associated with it. These materials could be polymers, metals, ceramics or composites.
Four basic criteria must be met for the self-healing strategy to be realized: storage, release, transport, and re-bonding. These criteria largely depend on the chemistry and properties of the healing agent system (Kessler 2007).
The microencapsulation of the healing agent is the first stage in satisfying the storage criteria. This is followed by dispersing the microcapsules and chemical trigger into the matrix well the material is being manufactured. It is worth pointing out that the healing agent should not react while it is being encapsulated and remain stable within the microencapsulate until is activated by the chemical trigger or catalyst embed with it.
Release is accomplished when the microcapsule wall is ruptured mechanically, such that the self healing agent is released. This implies that the microcapsules walls must be weak enough to break in the event o the structure being cracked.
In order for transportation to be successful solid or highly viscous liquids must not be used. It is important the surface energy of the transportation fluid to below relative to than of the fracture surfaces in order for it to penetrate the damaged area. It is also important for that, the healing agent is not too volatile as it must not evaporate or diffuse away from the damaged area before it is repaired.
Finally, once the healing agent is in the crack, it must Polymerization takes place when the healing agent reacts with the chemical trigger around the damaged area. This results in re-bonding of the cracked faces. The polymerization which occurs should be such that it is quick to take place at room temperature to avoid further damage.
Kessler goes further to state that the polymer produced should also have excellent adhesive strength with the fracture surface, the polymer film does not pull away from the crack faces that and finally, the polymerization should not require precise stoichiometric control and should be insensitive to non-ideal mixing with the initiator.
Research at the Autonomous Materials System at the University of Illinois demonstrated that healing properties can be placed in materials by embedding them in microcapsules. These properties are released to the defective areas upon fracture. They not only repair the damaged structure but also restore the mechanical performance.
Structural polymers are prone to cracks deep within the structure, which are difficult to detect and repair. If such was to occur in a micro electric component it could result in electrical failure. Thermal and mechanical fatigue induced micro cracking which is a major challenge with polymers. Self healing agents and a catalytic chemical trigger within the epoxy matrix can restore structural integrity in polymers after cracking occurs. The crack ruptures the healing agents which releases it in the cracked plane via capillary action. This is referred to as autonomic healing.
Kessler also reports number of other approaches are being taken to achieve self-healing in materials some these include; the use of shape memory alloy actuators to close and heal cracks, Others have investigated a self healing skin structure using an ultraviolet (UV)-curable epoxy for the liquid healing agent. If the skin is damaged, the UVcurable epoxy is released and is cured by ambient sunlight, materials that toughen in response to stress before a fracture event occurs. Stress-stimulated cross linking or polymerization reactions that increase the toughness of the material at the vicinity of a high stress concentration is referred to as mechanochemical activation and several strategies were recently suggested
Automonic Materials is a leading company in development and commercial application of self-healing materials. At present it offers technology to extend coating lifetimes. The company aims at developing self-healing properties in material that are not easily assessable to maintenance. It has a strong partnership with the University of Illinois's Autonomous Materials Systems.
The benefits of such these properties in materials would be reduced maintenance and inspection schedules as there would be longer lifespan and enduring strength. These materials could be applied in civil infrastructure, aerospace and electronics.
Self healing properties
Diagram 1
Application of these materials can be seen in aerospace (ref)
Self configuring
This is presently seen in self reconfiguring robots. These robots have shape changing abilities. They are able to accomplish this by rearranging existing connections between various parts. The rationale behind this is for the robot to be able to adapt to changing circumstances, to carry out new tasks or to carry out repairs on it. Self-reconfigurable robots inspired by existing models which occur naturally some of which include tissue-regeneration, epimorphosis, morphallaxis, diffusion reaction, potential fields, cellular automata, and others. Shen 2006 demonstrates the use of a 'US patented hormone-inspired model which mimics how cells adjust and adapt to changes which change their initial configuration, in the SuperBot. 'The same model can also control distributed and reliable self-reconfiguration, locomotion, manipulation, topology discovery, role negotiation, synchronization, role-based behaviour, and detecting and repairing damage'.
The SuperBot is good example of a self reconfigurable system. It was built at the USC/ISI Polymorphic Robotic Lab. The modules in the robot are such that they have the capability of connecting or disconnecting from other modules and also communication ability with other modules.
An important feature of these robots is the connectors. These are responsible for allowing the robots to assume various shapes and configurations. The connectors need to be such that they are flexible with the added ability to engage or disengage when one end of the connector is not operational.
Shen 2006 demonstrates the use of such connectors in what is referred to as the 'SINGO'.
These systems present a case for applications in aerospace and underwater in relation to self assembly systems and also in manufacturing in the development and use of tools and devices which are self reconfiguring to adapt to changing shop floor or process requirements.
Prognostics
Maintenance has been moving from 'fail and fix to 'predict and prevent'. This is evident with the use of maintenance techniques as RCM which embodies the use of CBM. Prognosis entails the ability to predict the remaining life an item or system. This prediction requires a high level of detail in terms of being precision and accuracy. The maintenance technique of CBM presents a challenge in terms of being used in prognosis. The technique is capable of telling us what is going on with the equipment being monitored, when faults develop but cannot accurately give the time period between fault initiation and failure. The emphasis in the application is fault detection and not monitoring degradation over time so as predict future behaviour of the system.
The following are the current approaches to prognosis ( Vachtsevanos et al 2006)
Current prognostic approaches include;
Model based; which is data collected from a model based simulation under normal and deteriorating conditions
Models based on random load conditions or modes
Data driven prognostic methods. This predicts the remaining useful life of items by monitoring the trajectory of a developing fault until it reaches a point requiring action.
Hybrid models; this is simply a combination of any of the above models
(ref)A number of systems have been developed with prognosis at the heart of it. The Joint Strike Fighter (JSF) has been designed with the capability to perform prognosis. When in use it monitors the condition of various components and sends this data to a station in real time where this data is processed. The data obtained gives a picture of the current state of the components the aim of which is to detect fault conditions before the event of failure.
The is done to reduce maintenance time and increase operational time, as the data received from the JSF gives information in regards of action to be carried out which in this instance is maintenance repairs or replacements. A vast number of applications in relation to prognosis application have been documented. (JSF) This demonstrates that information obtained via sensors in the aircraft can be used to increase availability, by using information on the current state of monitored items.
The information obtained from this system gives no indication on how long it will be before the item or system in general fails in other words it is more focused on failure prediction in addition to this it is application specific.
This observation and more is also capture by Lee et al 2006. He highlights that in spite of the recent developments in prognosis certain fundamental issues are yet to be addressed these include;
1. Most of the developed prognostics approaches are application or equipment specific. A generic and scalable prognostic methodology or toolbox doesn't exist;
2. Currently, methods are generally focused on solving the failure prediction problem. The tools for system performance assessment and degradation prediction has not been well addressed;
3. Features used for prognostics need to be further developed;
4. Many developed prediction algorithms have been demonstrated in a laboratory environment without industrial validations.
Intelligent prognosis is a systematic approach to continuously monitoring deterioration in a machine and estimating health indicators at a given instance of time to predict below performance overtime and at the same time determining which components in the machine are mostly likely to fail.
It is with this in mind that a computational prognostic algorithm with a software toolbox, developed by the Centre for IMS called Watchdog Agent has been developed to address these issues. The system has been shown to be capable for predicting degradation of items. It uses the readings from a number of sensors which measure critical properties of the item. It performs prognosis by trending and statistical modelling, allowing it to predict the future behaviour of data obtained via the sensors and as a consequence the behaviour of the item.
The software toolbox consists of a number of prognostics tools, and is based of open architecture. This allows interchangeably in the use of tools as a consequence of various applications.
Watchdog Agent™ demonstrates the existence of a prognostics tool which has the capability of predicting the significant remaining useful life of items by monitoring degradation and not faults detecting. This has been done without reliance on expert systems or external help and has been successfully applied in industrial test (Lee 2006)
E maintenance
The next generation maintenance system will have to be responsive to change (Labib 2008). Information would be a key to in order for maintenance systems to respond to change. The challenge however would be to have the right information in the right place at the right time. Information allow is not always sufficient in decision making knowledge and intelligence are. It can be inferred from the key features the unmet needs in current maintenance systems intelligence is going to play a major role in the next generation maintenance systems.
The use of the internet to provide real time data and analysis seems be capable of addressing these challenges. This brings us to the concept of e-maintenance.
There is not universal definition of e-maintenance. E maintenance is a concept which seeks to integrate ICT into the maintenance strategy and e business as a whole.
Muller et al 2006 gives various views on e-maintenance these include e-maintenance as;
Maintenance strategy
Maintenance plan
Maintenance type
Maintenance support
After reviewing these various views, a definition of e-maintenance was proposed which took into consideration the European Standard (EN 13306:2001) for maintenance terminology and at the same time examining e-maintenance as a component of e-manufacturing. The proposed definition goes thus;
'Maintenance support which includes the resources, services and management necessary to enable proactive decision process execution. This support includes e-technologies (i.e. ICT, Web-based, tether-free, wireless, infotronics technologies) but also, e-maintenance activities (operations or processes) such as e-monitoring, e-diagnosis, e-prognosis, etc.'
This definition captures the, Data to Information to Decisions, Off-line to On-Line to Real-time as key features of the next generation maintenance systems.
It can also be said e-maintenance deals with the challenge of synchronising various maintenance activities such the best decision on how activities can be prioritised, optimised and scheduled.
E-maintenance allows for better use of resources, activities as these can be integrated and synchronised. Activities such as fault diagnosis can be done from anyway in the world (remotely) as experts required to perform this task they can do it from where they provided they have assess to the right tools and information in there location. One distinct advantage is that it allows plants to be connected with expert centres. This has a huge number of benefits.
Remote monitoring of equipment can provide real time feedback to equipment designers to help them improve the future designs. This advantage of this would be data under exact operating conditions; this will also give the designers insight into the maintenance requirements and allow then to take this into consideration in design.
Lee et al 2006 makes the point that e-maintenance based on intelligent prognostics, where maintenance actions are synchronized with the overall operation of the system as well as with the necessary maintenance resources and spare parts. Such synchronization of maintenance actions and information flow infrastructure should enable autonomously triggering of services and ordering of spare parts, yielding near-zero downtime system operation through proactive, cost-effective maintenance that is the least intrusive on the normal function of the system
Conclusion and Recommendation
A literature review has been carried out on the next generation maintenance systems. Research areas and existing technology to satisfy the unmet needs have been identified.
One of the challenges of maintenance is the gap between technological advancements and maintenance practices. It must be said that most of the applications and research being carried out has seen much input from industry. The industry can be said to be the main stakeholder and end user of whatever technology is being developed. There needs to be a buy in from industry so as to reduce the gap between technological advancements and maintenance practice.
